Patent Application
20190143963
The invention relates to a method and a device for determining a safe speed at a future waypoint of a moving vehicle. Unless explicitly stated, the terms maximum coefficient of friction and coefficient of friction are used interchangeably below.
In principle, the assessment of the road conditions is up to the driver, who has to adapt his driving style to the former. Vehicle control systems such as ESC (Electronic Stability Control) or TCS (Traction Control System) or ABS (Antilock Braking System) help the driver to stabilize the vehicle in the limit range, supporting the driver in fulfilling the driving task in extreme situations. The effectiveness of such vehicle control systems essentially depends on the available maximum coefficient of friction μ (also referred to as maximum adhesion coefficient) at the current waypoint. There, the interaction between tire, surface and intermediate medium is crucial. Wet roads, snow and ice considerably reduce the available coefficient of friction between tires and road surface compared to the coefficient of friction available on a dry road surface. Suddenly changing coefficients of friction, such as those caused by changes in environmental conditions, can result in unstable driving situations and thus increase the risk of accidents. It is particularly dangerous if the driver of the vehicle approaches a curve too fast due to an incorrect assessment of the existing coefficient of friction.
Up to now, safe cornering speed is determined solely based on the map data of the routing of a road. Furthermore, a constant max. coefficient of friction (frequently p=1) is assumed. Ideally, the computation also includes a vehicle model that reflects the characteristics of the vehicle in question. In addition, for a known max. coefficient of friction, the vehicle can be decelerated at a distance from the curve to enable it to easily pass through.
The invention addresses the problem of providing a method which can be used to determine a safe speed at a future waypoint of a vehicle moving along a route with a known course. Another problem the invention addresses is providing a suitable device.
These problems are solved by a method and a device according to the features of the independent claims. Advantageous embodiments will be apparent from the dependent claims.
A method is proposed for determining a safe speed at a future waypoint of a vehicle, which moves along a route with a known course, and comprises the following steps: providing at least one item of route data characterizing the course of the route; providing a first probability distribution of a max. coefficient of friction at the current waypoint and/or at the future waypoint of the vehicle; determining a second probability distribution of a vehicle speed at the future waypoint from the at least one item of route data and the first probability distribution; and determining the safe speed from the second probability distribution.
The method according to the invention is based on the consideration of taking into account not only the route ahead of the vehicle, but also possibly changing environmental conditions, which can considerably limit the maximum forces that can be transmitted at the tire. The starting point is a continuous or discrete probability distribution of the maximum coefficient of friction present at the current waypoint of the vehicle and/or at the future waypoint of the vehicle.
The method can detect whether the vehicle is moving too fast for the prevailing ambient conditions. In the case of too high a speed, a need for action can be deduced therefrom before reaching the future waypoint, and be output e.g. in the form of data, warning or an automated driving intervention (vehicle deceleration, etc.). In automated vehicles, e.g. using one or more vehicle assistance systems, the safe speed may be used to compute a driving strategy, for instance by limiting an optimization space within which a velocity trajectory is sought, which in no way exceeds the safe speed for the future waypoint.
The safe speed at the future waypoint can be determined by selecting a q-quantile of the second probability distribution. As known to a person skilled in the art, a quantile is a measure of central tendency in statistics. The q-quantile corresponds to the integral from the chosen safe speed to infinity of the second probability distribution. This means that the actual speed at which the vehicle can pass the future waypoint in a stable manner is greater than or equal to the safe speed with a probability of the selected q-quantile×100%. The safe speed at the future waypoint can be determined using the equation
∫v
In equation (1) v* is the safe velocity, P(v) the second probability distribution, and qs the selected q-quantile. The q-quantile qs is given and the safe velocity v* is determined using equation (1).
Alternatively, the p-quantile of the second probability distribution may be used to determine a safe speed, i.e., the integral from 0 to the safe speed v*. The equation below applies to the selected p-quantile ps
∫0v
A curvature of the curve at and/or before the future waypoint of the vehicle can be processed as the at least one item of route data. If the curvature of the curve is known, the second probability distribution can be determined from the equation
αx2+κ2(s)*vx4−g2*μ2(s)=0 (2)
In the formula, ax is the longitudinal acceleration of the vehicle, K is the curvature of the curve, s is the path, g is the gravitational constant, vx is the speed of the vehicle in the direction of its longitudinal axis (vehicle longitudinal speed), p is the coefficient of friction. The given equation (2) is based on the assumption that the vehicle can be regarded as a simplified mass point. Of course, more complex vehicle models can be used to determine the second probability distribution.
According to a further embodiment, the step of determining the second probability distribution is conducted based on the assumption that the vehicle passes unaccelerated along the route, in particular through the future waypoint. As a result, the term αx2 can be ignored in equation (2) and a simple conversion of the probability distribution of the coefficient of friction μ at the relevant waypoint s to the probability distribution of the vehicle speed vx can be conducted.
In particular, a value is selected as the q-quantile, which depends on the vehicle type, the chassis type or the selected driving mode. The q-quantile characterizes a safe speed processing system. The safe speed can be selected differently, depending on the type of vehicle, for instance a sports car, a comfortable sedan, an off-road vehicle, etc. If a vehicle has a vehicle assistance system, which can be used to set different suspension modes, then for instance, an individual q-quantile that takes the driving characteristics of the vehicle into account can also be defined for different modes.
The q-quantile is preferably chosen such that the resulting selected safe speed is less than the true safe speed at the future selected waypoint. In that way it can be ensured that no dangerous situation for the vehicle results at the future waypoint.
According to a further expedient embodiment, a device for determining a safe speed at a future waypoint of a vehicle that moves along a route with a known course is proposed. The device comprises a first means for determining a second probability distribution of a vehicle speed at the future waypoint from at least one item of route data characterizing the course of the route and from a first probability distribution of a max. coefficient of friction at the current waypoint and/or at the future waypoint of the vehicle; and a second means for determining the safe speed from the second probability distribution. The first probability distribution can be determined by a computing unit of the vehicle and provided for further processing to determine the safe speed.
The device according to the invention has the same advantages as described above in conjunction with the method according to the invention.
The device may comprise further means for executing the method described.
Below, the invention is described in more detail with reference to an exemplary embodiment in the drawings. In the drawings:
In a second step S2, a first probability distribution Pges of a max. coefficient of friction μ at the current waypoint s and/or at the future waypoint s* of the vehicle is provided. The probability distribution Pges (μ) may be provided in discrete or continuous form. The manner in which such a probability distribution is determined is not the subject matter of this method.
In general, the max. coefficient of friction can be determined by direct or indirect methods. The determination of the max. coefficient of friction by direct methods is called effect-based and can be subdivided into direct, active and direct, passive methods. In a direct, active method, an active intervention in the driving dynamics of the vehicle is effected by braking and/or steering. In a direct, passive method, there is no active intervention in the driving dynamics of the vehicle. Instead, there is only an observation of effects of the coefficient of friction on the tire tread, the vehicle and such in the course of driving maneuvers of the vehicle, which the latter performs to achieve a predetermined navigation destination. To measure the effects of the coefficient of friction and to infer a max. coefficient of friction therefrom with sufficient certainty, the transmission of high forces at the tire is a prerequisite.
For indirect, cause-based methods, the max. coefficient of friction is determined based on parameters that affect it physically. These may be, for instance, a tread pattern, the rubber compound of a tire, its temperature, an inflation pressure, the road surface, its temperature, its condition (e.g., snowy or wet), etc.
The estimation of a coefficient of friction at a future waypoint of a vehicle may be described by way of example using a method comprising the following steps: a first set of parameters, which were or are determined for a current waypoint of the vehicle and which characterize the max. coefficient of friction at the current waypoint of the vehicle, are used to perform a prediction of a first probability distribution for the max. coefficient of friction at the current waypoint of the vehicle using a Bayesian network. Further, a second probability distribution for the max. coefficient of friction at the future waypoint of the vehicle is estimated from a second set of data about the future waypoint. Finally, a resulting, combined probability distribution is determined from the first and the second probability distribution. To be able to estimate the distribution of the coefficient of friction at the future waypoint in front of the vehicle, according to this exemplary procedure, the variables affecting the max. coefficient of friction both under the vehicle (i.e., at the current waypoint) and in front of the vehicle are processed. By making use of a large number of available data under and in front of the vehicle, the prediction quality of the coefficient of friction distribution at the future waypoint is high.
In principle, other procedures for providing a probability distribution of the max. coefficient of friction at the current waypoint and/or at the future waypoint of the vehicle can be processed within the framework of this method.
By means of an analytical relationship and the route data(s), in particular the curvature of the curve κ, the present probability distribution Pges(μ) at the future waypoint s* can be converted into a probability distribution Pv(v), where v represents the vehicle speed of the vehicle. This determination is made as step S3 in the diagram shown in
As a further step S4, the safe speed v* is determined from the now determined probability distribution of the velocity Pv(v) at the waypoint s*. This procedure will be described with reference to
The upper half of
The conversion of the probability distribution Pges(μ) into the probability distribution Pv(v) of the velocity at the future waypoint s* is conducted, for instance, on the basis of the differential equation according to equation (2). In this formula, vx represents the speed of the vehicle in the direction of its longitudinal axis (vehicle longitudinal speed), ax the acceleration of the vehicle in the direction of its longitudinal axis (vehicle longitudinal acceleration), κ the curvature of the curve, s the path, g the gravitational constant and μ the coefficient of friction. In processing equation (2), the vehicle shall be considered as a mass point for the sake of convenience. Another vehicle model can also be used, however. The curvature of the curve κ results from the above-mentioned map data for the relevant waypoint, in this case the chosen future waypoint s*. For the sake of simplicity, it is assumed that the vehicle passes through the considered waypoint s* unaccelerated which is why the term αx is omitted from equation (2). This can be used to determine the probability value of the speed for the different coefficients of friction according to the probability distribution shown in
The determination of the safe speed v* at the future waypoint s* (also referred to as the look-ahead point) is made by selecting a q-quantile qs of the determined probability distribution Pv(v) as a function of the speed v. In
According to equation (1), qs corresponds to the integral of the probability distribution Pv over the interval of the safe speed v* to infinity. This means that the actual speed at which the future waypoint s* can be passed in a stable manner is greater than or equal to the selected speed v* with a probability of qs×100%. The larger the selected qs, the safer the speed selection.
To determine the safe speed, a specific value for qs is specified for a vehicle type (for instance sports car, sedan or off-road vehicle) and/or depending on a chassis (sports suspension, comfort chassis or chassis selection selectable by mode selection). The q-quantile qs is selected permanently for a vehicle type and/or a chassis type/driving mode of the vehicle. The safe speed v* can then be calculated from the given q-quantile qs and equation (1). It is obvious that the selected safe speed v* does not correspond to the true safe speed, which is actually unknown. However, the safe speed v* is chosen by an appropriate selection of the q-quantile, which is likely to be lower than the true safe speed.
An action strategy can be derived from the now available safe speed v* for the future waypoint s*. If it is determined that the actual speed of the vehicle at the waypoint s* is greater than the determined safe speed v* or only slightly below the safe speed v*, then a need for action can be derived therefrom. Such a need for action may include data or a warning of the driver or, in the case of an existing driver assistance system, an automated vehicle deceleration. In autonomous vehicles, the safe speed v* can be used to limit an optimization space within which a suitable velocity trajectory is determined.
The present method is based on the consideration of predetermining a criticality of the driving situation at this waypoint for existing friction data about a future waypoint on the basis of the computed safe speed at this waypoint and on a speed prognosis based on the utilization of the current coefficient of friction of the driver taking into account route data. This can be used to derive the mentioned action recommendation, as, e.g., a preconditioning of a suspension system, a deceleration of the vehicle or a warning to the driver.
| Number | Date | Country | Kind |
|---|---|---|---|
| 10 2016 208 675.8 | May 2016 | DE | national |
This application is a National Stage application of PCT/EP2017/060111 filed Apr. 27, 2017, which claims priority from German patent application serial no. 10 2016 208 675.8 filed May 19, 2016.
| Filing Document | Filing Date | Country | Kind |
|---|---|---|---|
| PCT/EP2017/060111 | 4/27/2017 | WO | 00 |
